Scholarlink

Purpose <p>Magnetic resonance imaging (MRI) is essential for diagnosing lumbar foraminal stenosis (LFS). However, access remains limited in China due to uneven equipment distribution, high costs, and long waiting times. Therefore, this study developed a lightweight deep learning (DL) model using sagittal CT images to classify LFS severity as a potential clinical alternative where MRI is unavailable.</p> Methods <p>A retrospective study included 868 sagittal CT images from 177 patients (2016–2025). Data were split at the patient level into training (<i>n</i> = 125), validation (<i>n</i> = 31), and test sets (<i>n</i> = 21), with annotations, based on the Lee grading system, provided by two spine surgeons. Two DL models were developed: DL1 (EfficientNet-B0) and DL2 (MobileNetV3-Large-100), both of which incorporated a Faster R-CNN with a ResNet-50-based region-of-interest (ROI) detector. Diagnostic performance was benchmarked against spine surgeons with different levels of clinical experience.</p> Results <p>DL1 achieved 82.35% diagnostic accuracy (matching the senior spine surgeon’s 83.33%), with DL2 at 80.39% (mean 81.37%), both exceeding the junior spine surgeon’s 62.75%. DL1 demonstrated near-perfect diagnostic agreement with the senior spine surgeon, as validated by Cohen’s kappa analysis (κ = 0.815; 95% CI: 0.723–0.907), whereas DL2 showed substantial consistency (κ = 0.799; 95% CI: 0.703–0.895). Inter-model agreement yielded κ = 0.782 (95% CI: 0.682–0.882).</p> Conclusions <p>The DL models achieved a mean diagnostic accuracy of 81.37%, comparable to that of the senior spine surgeon (83.33%) in grading LFS severity on sagittal CT. However, given the limited sample size and absence of external validation, their applicability and generalisability to other populations and in multi-centre, large-scale datasets remain uncertain.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep learning-based lightweight model for automated lumbar foraminal stenosis classification: sagittal CT diagnostic performance compared to clinical subspecialists

  • 期刊论文

  • 发表日期：

    2025年8月23日

查看全文
  • Jia-Wei Huang,
  • Yu-Long Zhang,
  • Kai-Yu Li,
  • Hua-Lin Li,
  • Hao-Bo Ye,
  • Yu-Han Chen,
  • Xin-Xiao Lin,
  • Nai-Feng Tian

摘要

Purpose

Magnetic resonance imaging (MRI) is essential for diagnosing lumbar foraminal stenosis (LFS). However, access remains limited in China due to uneven equipment distribution, high costs, and long waiting times. Therefore, this study developed a lightweight deep learning (DL) model using sagittal CT images to classify LFS severity as a potential clinical alternative where MRI is unavailable.

Methods

A retrospective study included 868 sagittal CT images from 177 patients (2016–2025). Data were split at the patient level into training (n = 125), validation (n = 31), and test sets (n = 21), with annotations, based on the Lee grading system, provided by two spine surgeons. Two DL models were developed: DL1 (EfficientNet-B0) and DL2 (MobileNetV3-Large-100), both of which incorporated a Faster R-CNN with a ResNet-50-based region-of-interest (ROI) detector. Diagnostic performance was benchmarked against spine surgeons with different levels of clinical experience.

Results

DL1 achieved 82.35% diagnostic accuracy (matching the senior spine surgeon’s 83.33%), with DL2 at 80.39% (mean 81.37%), both exceeding the junior spine surgeon’s 62.75%. DL1 demonstrated near-perfect diagnostic agreement with the senior spine surgeon, as validated by Cohen’s kappa analysis (κ = 0.815; 95% CI: 0.723–0.907), whereas DL2 showed substantial consistency (κ = 0.799; 95% CI: 0.703–0.895). Inter-model agreement yielded κ = 0.782 (95% CI: 0.682–0.882).

Conclusions

The DL models achieved a mean diagnostic accuracy of 81.37%, comparable to that of the senior spine surgeon (83.33%) in grading LFS severity on sagittal CT. However, given the limited sample size and absence of external validation, their applicability and generalisability to other populations and in multi-centre, large-scale datasets remain uncertain.